Báo cáo y học: "Economic instruments for Obesity Prevention: Results of a Scoping Review and Modified Delphi Surve" pptx

This Provisional PDF corresponds to the article as it appeared upon acceptance. Fully formatted
PDF and full text (HTML) versions will be made available soon.
Economic instruments for Obesity Prevention: Results of a Scoping Review and
Modified Delphi Survey
International Journal of Behavioral Nutrition and Physical Activity 2011,
8:109 doi:10.1186/1479-5868-8-109
Guy EJ Faulkner ()
Paul Grootendorst ()
Van Hai Nguyen ()
Tatiana Andreyeva ()
Kelly Arbour-Nicitopoulos ()
M CHRISTOPHER Auld ()
Sean B Cash ()
John Cawley ()
Peter Donnelly ()
Adam Drewnowski ()
Laurette Dube ()
Roberta Ferrence ()
Ian Janssen ()
Jeffrey LaFrance ()
Darius Lakdawalla ()
Rena Mendelsen ()
Lisa M Powell ()
W BRUCE Traill ()
Frank Windmeijer ()
ISSN 1479-5868
Article type Research
Submission date 16 March 2011
Acceptance date 6 October 2011
Publication date 6 October 2011
Article URL />This peer-reviewed article was published immediately upon acceptance. It can be downloaded,

printed and distributed freely for any purposes (see copyright notice below).
International Journal of
Behavioral Nutrition and
Physical Activity
© 2011 Faulkner et al. ; licensee BioMed Central Ltd.
This is an open access article distributed under the terms of the Creative Commons Attribution License ( />which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Articles in IJBNPA are listed in PubMed and archived at PubMed Central.
For information about publishing your research in IJBNPA or any BioMed Central journal, go to
/>For information about other BioMed Central publications go to
/>International Journal of
Behavioral Nutrition and
Physical Activity
© 2011 Faulkner et al. ; licensee BioMed Central Ltd.
This is an open access article distributed under the terms of the Creative Commons Attribution License ( />which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
- 1 -
Economic instruments for Obesity Prevention: Results
of a Scoping Review and Modified Delphi Survey

Guy EJ Faulkner
1§
, Paul Grootendorst
2
, Van Hai Nguyen
2
, Tatiana Andreyeva
3
, Kelly
Arbour-Nicitopoulos
4
, M Christopher Auld

5
, Sean B. Cash
6
, John Cawley
7
, Peter
Donnelly
1
, Adam Drewnowski
8
, Laurette Dubé
9
, Roberta Ferrence
10
, Ian Janssen
11
,
Jeffrey LaFrance
12
, Darius Lakdawalla
13
, Rena Mendelsen
14
, Lisa M. Powell
15
, W.
Bruce Traill
16
, Frank Windmeijer
17

1.
Faculty of Physical Education and Health, University of Toronto, 55 Harbord
Street, Toronto ON, M5S 2W6, Canada
2. Leslie L. Dan Faculty of Pharmacy, University of Toronto, 144 College Street,
Toronto ON, M5S 3M2, Canada
3. Rudd Center for Food Policy & Obesity, Yale University, Department of
Psychology, 309 Edwards Street, New Haven CT, 06520-8369, USA
4. Department of Kinesiology, McMaster University, 1280 Main Street West,
Hamilton ON, L8S 4K1, Canada
5. Department of Economics, University of Victoria, PO Box 1700 STN CSC,
Victoria, BC, V8W 2Y2, Canada
6. Friedman School of Nutrition Science and Policy, Tufts University, 150
Harrison Avenue, Boston MA, 02111, USA
7. Departments of Policy Analysis and Management, and Economics, Cornell
University, 3M24 MVR Hall, Ithaca NY, 14853, USA
8. Center for Public Health Nutrition, School of Public Health, University of
Washington, Seattle, WA 98195-3410, USA
9. Desautels Faculty of Management, McGill University, 1001 Sherbrooke St
West, Montreal QC, H3A 1G5, Canada
10. Ontario Tobacco Research Unit, Dalla Lana School of Public Health,
University of Toronto, 33 Russell St., Toronto ON, M5S 2S1, Canada
11. School of Kinesiology and Health Studies Queen's University, 28 Division St.
Kingston ON, K7L 3N6, Canada
12. Department of Economics, Monash University, Building H4, Room 47
Sir John Monash Road, Caulfield, Victoria 3145, Australia
13. Schaeffer Center for Health Policy and Economics, University of Southern
California, 3335 S. Figueroa St, Unit A, Los Angeles, CA 90089-7273, USA
14. Ryerson University, School of Nutrition, 350 Victoria Street, Toronto ON,
M5B 2N8, Canada

15. Institute for Health Research and Policy University of Illinois at Chicago,
1747 W. Roosevelt Road, Chicago, IL 60608, USA
16. Department of Food Economics and Marketing, University of Reading
Whiteknights PO Box 237, Reading RG6 6AR, UK
- 2 -
17. Department of Economics, University of Bristol, 8 Woodland Road, Bristol
BS8 1TN, UK


§
Corresponding author: Guy Faulkner, Faculty of Physical Education and Health,
University of Toronto, 55 Harbord Street, Toronto, ON Canada M5S 2W6.

Email addresses:
GF:
PG:
VHN:
TA:
KAN:
CA:
SC:
JC:
PD:
AD:
LD:
RF:
IJ:
JL:
DL:
RM:

LP:
BT:
FW:
- 3 -
Abstract
Background
Comprehensive, multi-level approaches are required to address obesity. One
important target for intervention is the economic domain. The purpose of this study
was to synthesize existing evidence regarding the impact of economic policies
targeting obesity and its causal behaviours (diet, physical activity), and to make
specific recommendations for the Canadian context.
Methods
Arksey and O’Malley’s (2005) methodological framework for conducting scoping
reviews was adopted for this study and this consisted of two phases: 1) a structured
literature search and review, and 2) consultation with experts in the research field
through a Delphi survey and an in-person expert panel meeting in April 2010.
Results
Two key findings from the scoping review included 1) consistent evidence that weight
outcomes are responsive to food and beverage prices. The debate on the use of food
taxes and subsidies to address obesity should now shift to how best to address
practical issues in designing such policies; and 2) very few studies have examined the
impact of economic instruments to promote physical activity and clear policy
recommendations cannot be made at this time. Delphi survey findings emphasised
the relatively modest impact any specific economic instrument would have on obesity
independently. Based on empirical evidence and expert opinion, three
recommendations were supported. First, to create and implement an effective health
filter to review new and current agricultural polices to reduce the possibility that such
policies have a deleterious impact on population rates of obesity. Second, to
- 4 -
implement a caloric sweetened beverage tax. Third, to examine how to implement

fruit and vegetable subsidies targeted at children and low income households.
Conclusions
In terms of economic interventions, shifting from empirical evidence to policy
recommendation remains challenging. Overall, the evidence is not sufficiently strong
to provide clear policy direction. Additionally, the nature of the experiments needed
to provide definitive evidence supporting certain policy directions is likely to be
complex and potentially unfeasible. However, these are not reasons to take no action.
It is likely that policies need to be implemented in the face of an incomplete evidence
base.
- 5 -
Background
The causes of overweight and obesity, and the potential solutions to prevent
and reduce obesity prevalence are complex. We live in an obesogenic environment
that increasingly promotes a high energy intake and sedentary behaviour [1]. No
single strategy will solve this health problem. Social-ecological theory emphasizes
that physical activity and dietary behaviours are influenced by factors across multiple
domains including the individual, social, physical and policy spheres. Accordingly,
comprehensive, multi-level approaches are required to address obesity. One
important target for intervention is the economic domain.
Standard economic theory hypothesizes that individuals make decisions to
make themselves as well off as possible. In other words, individuals attempt to satisfy
objectives subject to constraints. Both objectives and constraints are germane to diet
and physical activity choices [2]. On the objectives side, economics emphasizes that
human welfare depends on multiple factors, and individuals make trade-offs between
them. If health were the only goal, then there likely would be little obesity and
individuals would spend all of their time and money on health-enhancing activities.
Clearly this is not the case. Obesity, then, could be the result of the trade-off that
individuals make between health and other desired goods, such as the consumption of
calorie-rich food and beverages, in order to maximize self-perceived welfare. While
preferences are certainly relevant, they alone cannot explain the dramatic increase in

obesity prevalence over the last several decades – it seems unlikely that preferences
for calorie-rich food or physical (in)activity have changed so suddenly. What have
changed are: the allocation of time, budget constraints, and technology. Changes in
diet and activity can be interpreted as optimizing responses to these changes. In
- 6 -
particular, the total (money plus time) price of consuming calorie dense food and
beverages has declined and this has at once reduced the price of calories and increased
purchasing power. At the same time, higher wage rates and longer hours spent in
sedentary employment have made physical activity more expensive. Standard
economic theory predicts that these price changes would rationally lead individuals to
increase caloric intake and reduce caloric expenditure. An important implication is
that, changing prices of calorie dense, unhealthy foods relative to that of low-energy,
healthy foods, or altering the cost of physical activity relative to that of sedentary
alternatives may lead to changes in diet and physical activity.
While taxes and subsidies are obvious candidates for a government to use to
change relative prices, the basis for the government to intervene on obesity is less so.
Early rationale for government intervention in obesity focuses on the negative externality
of obesity, which argues that obesity results in large health care costs and these costs are
borne collectively, so that obesity imposes financial externalities on those who are not
obese. While this argument has some merit, there are limitations. While the individuals
who are obese likely incur higher health care costs than those individuals who are not
obese at any given age, they also have shorter lifespans [3]. Hence, those individuals who
are obese might have lower total lifetime healthcare costs than those who are not obese.
Recent justification for government intervention rests upon insights from
behavioural economics. This literature suggests that the self-control problem could be
grounds for government intervention. These economists think of individuals as having
two ‘selves’: a relatively myopic ‘today’s’ self – which is the one that makes diet and
physical activity decisions – and a relatively far-sighted ‘future’ self, which lives with the
health consequences. There is sometimes a conflict between the two selves: Today’s self
may not adequately take into account future self’s welfare and succumb to the

temptations of calorie rich foods and sedentary lifestyles. The theory suggests that
- 7 -
individuals who recognize this dilemma – ‘sophisticates’ – will use self-commitment
devices (e.g., diets, fitness club memberships) to make today’s self account for the
consequences of their decisions on their future selves. Excise taxation of unhealthy foods
or physical activity subsidies can be thought of as a commitment devise to improve the
‘future-selves’ welfare of non-sophisticates [4,5].
It is worth noting that taxes and subsidies are not free tools. The public finance
literature documents a number of costs associated with the use of these tools. First,
consumers in the absence of taxes are maximising their utility. Taxes might distort
their choice and break this optimality and thus reduce consumer welfare and create a
deadweight loss. How large is this deadweight loss depends on elasticity of supply
and demand for a good. Second, there are labor and administrative costs required in
implementing the tax or subsidy policies. Third, there may be also budgetary costs,
especially in the case of subsidies. For example, policies that reward desired
behaviours, such as subsidies to physical activity, will create windfall gains to those who
already engage in the desired behaviour. This may make the policy a costly way to
change relatively few individuals’ behaviour.
Economic theory also informs us on potential challenges in applying the taxes
and subsidies to change individuals’ behaviours. First, in response to price changes,
individuals may substitute lower priced goods for higher priced ones. This is the
substitution effect. To illustrate, suppose that government decides to apply a special tax
on cola. Individuals might then substitute root beer or other kinds of sodas for cola. If
governments tax all sodas, then individuals might switch to sugar added sports drinks. As
a result of these potential substitutions, the weight of an individual can remain
unchanged. Second, taxes increase the prices, which in turn reduce the purchasing power
of one’s income. The reductions in purchasing power can affect diet and physical activity
choices in a way that can mute the effectiveness of tax policies in controlling weight. For
- 8 -
example, if people spend a lot of their budget on unhealthy food and the government

imposes an excise tax on these foods, people respond by reducing, but not eliminating,
their consumption of these foods. This reduces their purchasing power. In response,
individuals reduce their consumption of relatively expensive fruits and vegetables and
other healthy foods, and consequently, this leads to little change in weight.
The successful public health strategy of using tobacco taxation to reduce
smoking presents a strong case for considering an economic approach in the context
of obesity. Tobacco taxation has been recognized internationally as one of the most
effective population-based strategies for decreasing smoking prevalence and
consumption and the adverse health consequences [6]. Historically, the effectiveness
of taxation as a tobacco control measure has been evaluated in the context of price
elasticity of demand, the extent to which the consumption of a product (cigarettes)
falls or rises after a change in its price. Recent research estimates that, in high income
countries, a 10% increase in cigarette prices results in a 3% to 5% decrease in demand
for cigarettes among adults [7]. While price elasticity estimates are comparable
among high income countries [8], the impact of taxation appears to be greater in low
and middle income countries, where smoking rates are generally higher and tobacco
control policies weaker [9].
Existing reviews examining the effects of economic incentives or
disincentives on food consumption, physical activity and/or obesity, including a
recent brief Canadian parliamentary report [10], conclude that little is known about
the impact of various economic instruments on healthy eating and physical activity or
on their effectiveness in preventing and controlling obesity in general, or that the
impact will be modest [11-14]. However, despite an incomplete evidence base, policy
makers still need to select and implement interventions that may have a significant
- 9 -
population health impact. Given that the current situation is described as ‘a
cacophony of policy in which different analyses and policy solutions have been
developed and proffered, each clamouring for support, funding and adoption’[15],
there is clearly an urgent need for clear and solid evidence to emerge and be
synthesized to guide policy decisions.

The purpose of this project was to synthesize existing evidence regarding the
impact of economic policies targeting obesity and its causal behaviours (diet, physical
activity), and to make specific recommendations for the Canadian context. To
achieve this, we adopted Arksey and O’Malley’s [16] methodological framework for
conducting scoping reviews. Scoping reviews are distinct from systematic reviews in
that a) they often address broad topics where a variety of study designs and secondary
topics may be relevant, b) they are less likely to formally assess the quality of
included studies or use study quality criteria to guide the synthesis of data, and c) they
are used to identify parameters around a body of literature, and to identify gaps in the
existing body of research. This study’s scoping review consisted of two main phases:
1) a structured literature search and review, and 2) consultation with experts in the
research field through a Delphi survey and workshop. The former phase focused on
selecting and reviewing empirical studies that look at weight outcome and use
population level survey data. Given that our interest is in the application of
population-level intervention such as taxes and subsidies and given the concern about
the substitution effects (discussed earlier), we believe that the evidence from studies
of this type are the most relevant. Our study selection thus distinguishes our review
from recent reviews such as that of Thow and colleagues [17] which covered both
empirical and simulation studies that estimated the effects of subsidies and taxes on
specific food products on consumption, body weight and chronic conditions. As a
- 10 -
point of comparison, Thow et al. [17] included a total of 24 studies but only 6 of these
investigated body weight using population survey data, while our review assesses 20
studies of this type. The latter expert consultation phase was a critical component of
the review. As the study of economic interventions for obesity is a relatively nascent
area of inquiry, much of the knowledge may not appear in the published or grey
literature. Soliciting input from experts regarding an incomplete body of evidence
was deemed a necessary step.

Structured Literature Search and Review

The search strategy was designed in consultation with an information
coordinator from the Ontario Tobacco Research Unit (OTRU). Detailed and extensive
searches were conducted on Medline, PsycInfo, PubMed, Econlit, Policyfile, Pais
International, OVID, Web of Science, Cochrane Reviews and Google Scholar from
September to December 2009. Searches were also performed on a range of grey
literature sources including NBER, U.S. Department of Agriculture (USDA)’s
Economic Research Service, AgEcon Search, and other governmental agency
websites. Search term combinations were used to identify relevant studies in the
nutrition domain: ‘overweight, fat, diet, nutrition, caloric, weight, obesity, BMI,
consumption, demand, intake’ with ‘taxes, subsidy, intervention, economic policy,
transfer program, income support, WIC, food stamp, cash transfer, agriculture
subsidies, farm policy’. Reviews were done of retrieved primary and review article
reference lists (including those from previously published systematic reviews); hand
searches of key nutrition and health economics journals (to December, 2009), and;
expert panel members were asked to review the final reference list for completeness.
- 11 -
In May 2010, an updated literature search was completed that focused from December
2009 to May 2010.
This comprehensive search resulted in 1198 potentially relevant studies. The
initial screening identified 379 studies that employed empirical analysis. Next, studies
were selected that focused on financial measures such as prices, subsidies, taxes or
income transfer programs as the central intervention. Given the availability of an
existing comprehensive review [18] and limited time, we excluded studies that
focused on the effects of food prices on food consumption and demand. Instead,
studies that explicitly focused on weight outcomes (such as obesity and body mass
index), physical activity or caloric intake were assessed. Finally, the review focused
solely on observational or randomized controlled trial (RCT) studies that estimated
behavioural responses and hence, excluded simulation studies. These requirements
further reduced the number of studies reviewed to 38 (see Figure 1) which included
20 studies assessing tax or food subsidies, 4 studies assessing agricultural policies and

subsidies, 4 studies assessing physical activity outcomes (tax credits, gas prices, road
congestion taxes), and 10 studies assessing targeted income transfer programs. In
addition to these empirical studies, we identified 7 relevant reviews. Data from each
empirical study was abstracted (e.g., authors, study location, year of publication)
including some analytical detail such as design and key findings. Next, the abstracted
information was collated and summarized in chart form.
INSERT FIGURE 1 HERE
Expert Panel
To supplement the literature review, an expert panel was convened to
contribute input to the literature search strategy, identification of grey literature, and
then to assess the strengths and limitations of different economic approaches with a
- 12 -
view to proposing specific recommendations for the Canadian context. Expert
consultation is an increasingly acceptable source for gathering evidence about a topic
particularly when the extant literature is weak [19]. Panel members were identified
through an initial literature search for researchers prominent in the field of economics
and obesity. Each panel member was also asked to nominate others who should be on
the panel. We sought to recruit a varied panel of experts who had published in the
area of economic instruments and obesity and were familiar with the literature
concerning a broad range of economic instruments. The primary inclusion criteria
included publications on economic interventions or consequences of obesity as
indicative of specialist knowledge. Turoff [20] recommends panels between 1 and 50.
Based on project time lines and cost considerations, we recruited 12 experts from
Europe (n=2), Canada (n=3) and the United States (n = 7).
We adopted a Delphi survey approach in consulting with the experts. The
Delphi survey is a mixed-method research approach, designed for exploring the range
of opinions, and exploring (or achieving) consensus on a specific topic. The
technique is considered particularly useful in areas of limited research or in areas
where there is controversy, debate or lack of clarity [21]. The Delphi has been
successfully applied to a range of issues, including views on the most suitable

monetary incentives on food to stimulate healthy eating [22]. A four-round
conference style format was applied in the current project to examine economic
instruments for addressing obesity. During round one, telephone calls were conducted
with all panel members to identify potential economic instruments and to confirm the
project’s literature search strategy. Based on these discussions and the reviewed
literature, a survey was created that listed the most commonly reported economic
instruments (see Table 1). In round two, this survey was sent to all participants with
- 13 -
the request to rate each instrument in terms of its potential impact on obesity,
consumption, its cost-effectiveness, potential for unintended benefits or harm,
equitability, and political feasibility (see [2]). Experts were then asked to return their
responses to a nominated facilitator external to the Delphi process. Respondent names
were removed, replaced with a number and then forwarded to the first author. During
the third round, questionnaires were returned to each individual expert, containing a
summary of their score for each item, along with the score for the group as a whole.
Panel members were invited to review their individual ratings against these group
means, and resubmit their responses with changed or unchanged scores. Final
responses were returned to the facilitator and forwarded to the lead expert. Group
means were calculated for each item and then ranked according to their score within
each major type of economic instrument. This ranking represented the group’s
consensus, and was distributed via email to the expert panel with summary statistics
purposely timed to precede a 1.5 day long in-person panel meeting held in Toronto.
At this meeting, panel member opinions and views on their Delphi rankings and
attendant recommended policies were solicited.
Results
Delphi Survey Results
Table 1 presents the findings of the final Delphi survey completed before the
panel meeting. First, all economic instruments were rated as having a relatively
modest impact, if any, on obesity. Economic instruments targeting consumption were
rated higher than those targeting physical activity. Of the instruments, changing

agricultural subsidies was rated as having the highest potential impact on obesity but
also the lowest in terms of feasibility. Food taxes were rated second highest in terms
of obesity impact but also scored highest in terms of potential for unintended harm
- 14 -
and being inequitable. Fruit and vegetable subsidies and beverage taxes were rated
similarly in terms of potential impact on consumption and obesity but differed in
terms of potential for unintended benefit and cost-effectiveness.
INSERT TABLE 1 HERE
The panel meeting discussions largely were in line with these survey findings
and concentrated on three broad considerations: 1) reviewing agricultural policy and
subsidies; 2) implementing a tax on caloric sweetened beverages; and 3) examining
how fruit and vegetable subsidies can be targeted. Attention now turns to discussing
each of these after a brief overview of the research evidence. The consideration of
food taxes, income transfers, and economic instruments for promoting physical
activity is briefly noted.
Reviewing Agricultural Policy and Subsidies
Evidence
Several authors and commentators [23-26] have taken the strong correlation
between increased farm subsidies and the rise in obesity rates in the US since the
1970s as evidence that they are causally related. They argue that these subsidies have
reduced the prices of soybeans, corn and other farm commodities. These subsidized
commodities – which are cheap sources of sugar and fat for processed foods – have
lowered the price of processed foods. The lower prices of these processed foods in
turn have contributed to their over-consumption, which has contributed to the increase
in obesity. Additionally, subsidies can provide incentives for technological
innovations which have increased the availability of ingredients such as corn-based
sweeteners.
- 15 -
Several empirical studies have challenged the claim that farm subsidies have
increased rates of obesity. Miller and Coble [27] investigated whether farm subsidies

make retail food products in the US more affordable using annual time series data
from Economic Research Service of USDA for the period 1961-2002. The
affordability of food, their outcome variable, is captured by the proportion of
disposable income spent on food while farm subsidies are measured by direct
payments to farmers. In addition to farm-to-retail price spread and consumer income,
their model also includes agriculture’s total factor productivity (TFP) to shed light on
the effect of changes in technology on food affordability. They estimated this
regression in the aggregate as well as across 6 specific food groups. The results
indicated that direct payment impact on food affordability was not statistically
significant. In contrast, the positive and statistically significant TFP implies that
advances in agriculture technology have increased the affordability of foods.
Furthermore, these results are consistent across food groups. These findings [27]
provide empirical evidence that cheap food prices are mainly caused by increases in
agricultural efficiency, perhaps enhanced by public R&D subsidies, over the last
several decades rather than by farm subsidies.
Beghin and Jensen [28] used historical data to examine whether US farm
policies for sweetener crops have affected the consumption and composition of
sweeteners in the US diet. The data showed that commodity programs have raised the
price of cane sugar and decreased the price of corn. At the same time, agricultural
R&D expenditure lowered the cost of corn more than that of sugar. Thus high fructose
corn syrup became an inexpensive substitute for sugar in food and beverages.
However, they emphasize that the effect of policy on ingredient prices has become
less important over time, with the current farm value share in sweetened food being
- 16 -
below 5%. They also noted that increased consumption of sweetened foods and
beverages are observed in other countries which have different or no commodity
programs.
Alston and colleagues [29] examined US and international data to shed light on
the impact of farm subsidies on commodity prices and of commodity prices on food
retail prices. They found that farm subsidies have had very modest (and mixed)

effects on the total availability and prices of farm commodities that are the most
important ingredients in more-fattening foods. Second, such small commodity price
impacts would imply very small effects on costs of food at retail, which, even if fully
passed on to consumers, would mean very small changes in prices faced by
consumers.
Additional evidence suggests that agricultural R&D subsidies contribute to high
productivity and thus reduce the prices of commodities. Alston et al. [30] studied the
contributions of US public agricultural research and extension investments over 1890-
2002 to state-specific agricultural productivity for the period 1949-2002. They found
that both state and federal agricultural R&D investment yielded high returns.
Specifically, a one dollar increment in investments in agricultural research and
extension by 48 U.S. states generated own-state benefits of between $2 and $58 and
averaged $21 across the states. They suggested that the returns would be even higher
if the spill-over effects across the states were taken into account (between $10 and
$70 per research dollar across the states, with an average of $32).
Panel Conclusions
Panelists agreed that agricultural support policies were influential in the rise in
rates of overweight and obesity. However, panelists conceded that there is no strong
- 17 -
empirical evidence linking agricultural support policies to the growth in obesity.
Nevertheless, panelists felt strongly that the lack of evidence was a result of the
complex causal pathway and considerable time lags between policy changes and
resulting changes in population levels of obesity. While panel members felt that
modifying agricultural policies would have the biggest impact on reducing obesity,
such modification also scored the lowest for feasibility. Limiting or eliminating farm
subsidies to commodity farmers is unlikely to rapidly change a complex agricultural
system that has evolved over decades. More attention could be directed toward the
impact of agricultural R&D supports which may have over time lowered cost of
added sugars and fats derived from corn, soy, potatoes and other farm commodities.
In turn, these added sugars and fats have found their way into much processed food. It

was also suggested that subsidies may have an impact on food formulation rather than
just price alone. Holding consumer food price equal, a subsidy that (for example)
increases the affordability of sweeteners could lead to undesirable dietary outcomes
without affecting product price.
Agricultural subsidy-specific recommendations included:
a) Create and implement an effective health filter to review new and current
agricultural polices to reduce the likelihood that such policies have a
deleterious impact on population rates of obesity. Specifically, an agricultural
support policy should become secondary to a food and health policy.
b) Restructure R&D investment and subsidies to promote increased development
of fruit and vegetable production and distribution. Measures to raise domestic
supply of fruits and vegetables can also be complemented by lowering tariffs
on imported fruits and vegetables.
c) Develop transportation and subsidized revenue insurance policies to assist
farmers who grow fruits and vegetables, widely considered a riskier crop than
other agricultural products. Farmers should be engaged as ‘anti-obesity’
partners.
d) Develop measures that promote easy access to fruit and vegetables for
Canadian households. For example, both the European Union and the United
States have recently implemented policies to actively promote farmers’
markets.
- 18 -
Caloric Sweetened Beverage Tax
Evidence
We identified 5 recent empirical studies that evaluated the effect of beverage
taxes currently adopted in a number of US states on body mass index (BMI) and
obesity [31-35]. The focus of these studies on the effect of beverage taxes on weight
outcomes differs from that of the beverage price elasticity studies, such as those
reviewed in [18], which estimated the effects of beverage prices on beverage
consumption.

All of these 5 studies used cross-sectional data. Two studies examined the
effect of soda taxes on adult weight. Kim and Kawachi [31] investigated the effect of
these taxes on state-level obesity prevalence, using state-level taxes on soft drinks and
snacks and the 1991-1998 cross-sections of data from the U.S. Behavioral Risk Factor
Surveillance System (BRFSS). They found weak statistical evidence (p-value = 0.09)
that states that had repealed a soft-drink or snack-food tax were 13 times more likely
than states with a tax to experience a relative increase in obesity prevalence. In
addition, states without a soft drink or snack food tax were four times more likely
(albeit statistically insignificant p-value=0.25) than states with a tax to exhibit a high
relative increase in obesity prevalence.
Fletcher et al. [32] also used the BRFSS data, but for the period 1990-2006,
and focused on the effect of soft drink taxes on individual weight outcomes. Using
their preferred model, they found that soft drink taxes influence individuals’ BMI but
the impact was small in magnitude. The authors also reported that tax effects on
weight outcomes were larger among low income groups. This result suggests that
such taxes may not be regressive as is commonly assumed.
- 19 -
Powell and colleagues [33] examined the effect of soft drink taxes on
adolescent weight using Monitoring the Future data for the period 1997-2006. They
found no statistically significant relationship between soda taxes and adolescent
weight outcomes but did find a weak economic and statistically significant
relationship between the vending machine soda tax rate and BMI among adolescents
at risk for overweight.
Fletcher et al. [34] considered the effect of soda taxes on children and
adolescent weight using the U.S. National Health and Nutrition Examination Survey
(NHANES). In addition to examining weight, they assessed whether higher soda
taxes lead to increased consumption of milk and juice. Their econometric framework
controls for unobserved state-level characteristics (through the inclusion of state-
specific fixed effects) that could be correlated with soft drink taxes (failure to control
for unobserved characteristics could lead to misleading estimates). They found that

soft drink taxes lead to a modest reduction in soda consumption by children and
adolescents, but have no effect on children and adolescents’ net weight due to an
increase in consumption of whole milk (but not juice or juice-related drinks). They
concluded soda taxes, as currently practiced, do not reduce weight in children and
adolescents. However, as children and adolescents appear to substitute whole milk for
soft drinks in response to soda taxes, there may be unexplored broader nutrient
benefits of soda taxes for children and adolescents.
Finally, Sturm et al. [35] investigated the effect of state sales taxes for soft
drinks on children’s consumption of soft drinks and weight gain. They used the tax
rates that were in effect in January 2004 and matched them to the fifth-grade wave of
the Early Childhood Longitudinal Study individual-level data. Their results indicated
that existing taxes on soft drinks do not substantially affect overall levels of soda
- 20 -
consumption

or obesity rates. However, they found that subgroups of at-risk

children
(i.e. children who are already overweight, come from

low-income families, or are
African American) may be more

sensitive than others to soft drink taxes, especially
when soft drinks are available at schools. They suggested that a larger soft-drink tax is
required to generate meaningful changes in consumption and weight outcomes.
Panel Conclusions
Three-quarters of the panel recommended moving forward with a tax on caloric
sweetened beverages; these include soda, energy drinks, sports beverages and many
juices and iced teas that are sweetened with sugar, corn syrup, or other caloric

sweeteners. Sugar-free diet drinks, diet beverages, sugar free juice, and flavoured
milk would be tax exempt. There was agreement that while such a tax may in itself
have a modest impact on obesity, it could be quite powerful in its impact over time,
and have a synergistic effect with other tax, legislative, and educational initiatives to
address obesity – this may be largely in relation to changing norms about dietary
consumption.
In some respects, such a recommendation was described as a “leap of faith”
given the incomplete evidence base. Nevertheless, most panelists felt that a tax on
caloric sweetened beverages is justified, for several reasons. First, it was noted that
unlike fast foods, caloric sweetened beverages “served no nutritional value”. Second,
there was no indication from the empirical evidence that such a tax would be
regressive and unfairly penalize low income individuals and households. Finally, just
as the (rather successful) tobacco control policies were introduced with imperfect
information concerning their effectiveness, so too, many obesity control policies will
need to be introduced in the context of imperfect evidence. The actual impact of
public policies will only be clear once they take effect.
- 21 -
The uncertainty surrounding the impact of such taxes is likely due to the low
level of existing tax rates since the current caloric-sweetened beverage taxes were not
designed and implemented to address obesity. These taxes are too low to generate a
meaningful influence on caloric consumption and body weight, and lack sufficient
variability across jurisdictions and time to generate precise causal effect estimates.
Even so, there is evidence that low income individuals and children are the most
sensitive to changes in food prices generally. Additionally, there is strong empirical
evidence that the consumption of soft drinks is responsive to its prices (for every 10
% rise in prices, consumption declines 8 to 10%; see [18]). Therefore, it is reasonable
to expect that a sufficiently high tax imposed on caloric sweetened beverages would
be likely to reduce consumption. However, experts acknowledged that such data
relies on many assumptions that may not hold during implementation. Unknown
effects regarding substitution, compensatory behaviour, and producer response all

serve to lower confidence in such claims.
In terms of magnitude, small taxes on soft drinks will likely do little to lessen
soft drink consumption or prevent childhood obesity. The panel experts noted that the
US tax rates up to 7.25% at the time did not have the desired effect on weight
outcomes. The level of the taxes will depend on where the soft drinks are sold
(vending machine, convenience stores, and supermarkets) but experts suggest a
minimum tax of around 20% of the price. For example, New York State was
considering an 18% tax in 2009. Further, taxes should be calculated and implemented
on a unit basis, rather than percent of price to avoid quantity discounts. Panellists also
suggested that the taxes be salient. That is, taxes are likely to have larger impacts if
they are made visible to consumers [36].
- 22 -
The proposed soft drink tax would also deliver several other benefits. First, the
revenue from this tax could be used to fund other initiatives to reduce obesity, such as
the introduction of free water fountains at public places. Tax revenue could also be
used to fund fruit and vegetable subsidy programmes. Second, a decline in soft drink
consumption will decrease sugar intake, which may provide health benefits beyond
the direct effects on obesity (see [37] for the list of cardiovascular diseases associated
with sugar consumption). Third, experts noted that fast foods and soft drinks are often
consumed together, so they may be complementary goods. Thus, higher soft drink
prices may reduce the consumption of fast foods.
Three panel members did not believe such a tax would have an impact on
obesity prevalence, claiming that consumers would substitute unhealthy foods not
subject to the tax if faced with higher prices for caloric sweetened beverages. They
favoured addressing the root economic forces at play, which they viewed as the
agricultural R&D and commodity subsidy policies that have lowered the prices and
increased the consumption of energy dense foods. An alternative strategy was taxing
sugar as an input
to caloric sweetened drinks rather than consumer-facing taxes. This
might encourage health-promoting reformulations as well as put some upward

pressure on consumer prices on such beverages.
Specific recommendations for the Canadian context included:
a) Apply the tax on the amount of caloric sweetener in the beverage (e.g., 10
cents per ounce of sweetener);
b) Rationalize the tax in terms of broader health benefits as opposed to a single
focus on obesity. Decreasing sugar consumption has health benefits beyond
those associated with lowering obesity;
c) Combine the implementation of such a tax with targeted unsweetened
beverages and/or fruit and vegetable subsidies, or in other obesity prevention
interventions, and;
d) Monitor any unintended consequences of the tax implementation in terms of
producers’ formulation responses[0].
Food taxes
- 23 -
Another policy option is to tax certain foods linked to obesity. The panel
viewed the evidential base for policy in this area to be more compelling than the
evidence to support a beverage tax. In particular, they were swayed by the 4
longitudinal studies that suggest that low fast-food prices increase weight outcomes
[38-41] and 6 studies employing cross-sectional data [42-47] . However, panel
members did not recommend proposing such taxes at this time despite the empirical
evidence. There are a number of difficulties with the design and implementation of
food taxes that require further research before specific recommendations can be made.
The greatest challenge with food taxes is defining the scope of foods to be
taxed that fall under the category of energy dense, unhealthy foods. Complex decision
rules to capture all foods is likely to be unfeasible and involve large administrative
costs which might even exceed the revenue from taxing such foods. A more narrow
scope may not achieve the goal because consumers will be able to substitute from one
energy-dense food that is taxed to some other energy dense food that is not taxed.
This could be averted by defining a tax on food types as proportional to its content of
sugar for example. However, deciding on what is actually being taxed is not without

controversy.
Another important concern with food taxes is the issue of food insecurity. For
low income individuals, cheap, high energy foods may be the primary source of
energy. Accordingly, these individuals may use a higher share of their income to pay
food taxes than their wealthier counterparts. In other words, food taxes are more
likely to be regressive. An additional challenge is that some foods have a mixture of
good and bad nutrients, such as cereals with added sugar. Taxing these foods might
eliminate both good and bad sources of nutrients which in turn might have a
deleterious impact on health.